A heuristic intelligent detection algorithm for reflection based on terahertz time-domain spectroscopy

In this paper, we utilize a terahertz (THz) reflection detection structure to examine multilayer heterogeneous samples and propose an improved non-destructive testing algorithm. Firstly, the Rouard equation is introduced to establish a THz wave reflection model. To enhance the model's accuracy, the Debye model is employed to calculate the dielectric constant, determining the refractive index of the samples. The model parameters are solved using genetic algorithms and the African vulture optimization algorithm. Experiments were conducted using a six-axis robotic arm and a THz spectrometer to detect non-metallic composite material samples. The results demonstrate that the proposed algorithm can accurately detect the samples, with a thickness measurement precision of 1 micrometer and an error margin within 4%. This study proposes a more reliable non-destructive testing algorithm based on terahertz time-domain spectroscopy technology, advancing the application of terahertz technology in the power industry and industrial production.